FortLP

Linear and mixed integer linear programming problems

FortLP was previously supported by NAG, but is no more. This page has been unlinked.

The present version of FortLP uses a sparse implementation of the revised simplex method for linear programming. A branch-and-bound technique is used to obtain integer solutions. The basis matrix is maintained in sparse LU form and is updated by the Forrest-Tomlin update procedure.

The software is written in ANSI standard Fortran 77. The current version uses INTEGER*2 constructs to preserve space, but these are likely to be removed at the next release, in light of increasing availability of storage. These nonstandard constructs may impose some restrictions on portability; otherwise, FortLP is expected to run on most machines. The package is being distributed as source code. It is fully supported and maintained, and regular updates are produced.

FortLP accepts problem data in MPSX format but is written in a modular way to allow users to use their own methods for supplying input.

Interior-point methods with a crossover to the simplex method during a final phase are being developed.

Need more info?

Visit the NAG WWW site or contact:

NAG, Inc.                  
1400 Opus Place, Suite 200 
Downers Grove, IL  60515-5702 
USA 
Phone: (708) 971-2337 
Fax: (708) 971-2706 

[ Optimization Software Guide | OTC Home Page | NEOS Server | NEOS Guide ]